After the opening of the electricity retail market

large power users are faced with the problem of how to find a suitable trading partner from a large number of power producers. As a service oriented enterprise

power grid companies have a good will to providerecommendation service for power consumers. In order to meet this new demand

a power purchase recommendation algorithm based on user collaborative filtering is proposed. In order to eliminate the unreasonable similarity calculation caused by the large difference of attribute elements in the data set

a new method based on Mahalanobis distance is proposed to form a similar user set. Furthermore

the proportion of purchasing power amount that represents the user’s preference for power plant is defined

which is used as the order basis for the recommendation of purchasing trading partners. Case studies based on Shandong electric power market transaction data are carried out. The results show that the algorithm is feasible.